Modal split - NPTEL

1 , travel demandmodelingis O-Dmatrixobtainedfromthetripdistribution is slicedinto number of matricesrepresentingeach canceandfactorsa ectingmode choiceproblemwillbe briefdiscussionontheclassi cationof mode choicewillbe types of mode choicemodelswillbe discussedin transportmode is probablyoneof themostimportant classicmodelsin becauseof thekey roleplayed by publictransportin useof roadspacemoree moresocialbene tslike if morepeoplebegintousepublictransport, therewillbe lesscongestionontheroadsandtheaccidents willbe publictransport,we cantravel withlow addition,thefuelis usedmoree ciently. Maincharacteristicsofpublictransportis thattheywillhave someparticularschedule, ,privatetransportis highly providesmorecomfortableandconvenient hasbetteraccessibility modechoice,therefore,is probablythesinglemostimportantelement in a ectsthegenerale ciencywithwhich we cantravel is important thento developandusemodelswhich aresensitive to thosetravel attributesthatin uenceindividualchoicesof uencingthechoiceof modeThefactorsmay be : Thefollowingfeaturesarefoundto be important:(a)caravailability and/orownership;(b)possessionof a drivinglicense;(c)householdstructure(you ngcouple,couplewithchildren,retiredpeopl eetc.)

2 ;(d)income;Introductionto V , 2007(e)decisionsmadeelsewhere,forexample theneedto usea carat work,take childrento school,etc;(f) : Modechoiceis stronglyin uencedby:(a)Thetrippurpose;forexample,th ejourneyto workis normallyeasierto undertake by publictransportthanotherjourneysbecauseo f itsregularity andtheadjustment possiblein thelongrun;(b)Timeof theday whenthejourneyis undertaken.(c)Latetripsaremoredi cultto accommodateby : Therearetwo types of quantitative andtheotheris qualitative. Quantitative factorsare:(a)relative travel time:in-vehicle,waitingandwalkingtimesby each mode;(b)relative monetarycosts(fares,fuelanddirectcosts); (c)availability andcostof parkingQualitative factorswhich arelesseasyto measureare:(a)comfortandconvenience(b)re liability andregularity(c)protection,securityA good modechoiceshouldincludethemostimportant of of , theobjective of transportationplanningwas to forecastthegrowthin demandforcartripssothatinvestment couldbe plannedto to be themostimportant determinants of modechoice,attemptsweremadeto a modelis thisway di erent characteristicsofthepersoncouldbe preservedandusedto thistimerelatedthechoiceof modeonlyto featureslike income,residentialdensity thatthesemodelscouldbe veryaccuratein theshortrun,if publictransportis availableandthereis thattheyareinsensitive to policydecisionsexample:Improvingpublictr ansport, no e ecton modalsplitaccordingto thepost-distributionmodel.

3 Thatis modalsplitis is possibleto includethecharacteristicsof thejourneyandthatof thealternative modesavailableto undertake is alsopossibleto bene V , calleddisaggregateif theyarebasedonhouseholdor thesimplestformof modechoice,wherethetravel choicebetweentwo modesis of each theutility of onemode is higherthantheother,thenthatmode is transportation,we have disutility hereis representedascij=a1tvij+a2twij+a3ttij+a4 tnij+a5 Fij+a6 j+ ( )wheretvijis thein-vehicletravel timebetweeniandj,twijis thewalkingtimeto andfromstops,ttijis thewaitingtimeat stops,Fijis thefarechargedto travel betweeniandj, jis theparkingcost,and is a thetravel costis low,thenthatmodehasmoreprobability of two modes(m=1,2)thentheproportionof tripsby mode1 fromzoneito zonejis(P1ij)Letc1ijbe thecostof travelingfromzoneito zonejusingthemode1, andc2ijbe thecostof travelingfromzoneito zonejby mode2,therearethreecases:1. ifc2ij-c1ijis positive, thenmode1 is ifc2ij-c1ijis negative, thenmode2 is ifc2ij-c1ij= 0 , thenbothmodeshave normallyexpressedby a logitcurve as shownin gure9:1 Thereforetheproportionoftripsby mode 1 is given byP1ij=T1ij=Tij=e c1ije c1ij+e c2ij( )Thisfunctionalformis calledlogit,wherecijis calledthegeneralizedcostand is gure9:1showstheproportionof tripsby mode1 (T1ij=Tij) againstcostdi of tripsfromzoneito zonejis 5000,andtwo modesareavailablewhich hasthecharacteristicsgiven in Table9 modebus,andthefarethatis collectedfromthemode thebusis reducedto 6, then given travel by car(Equation)=ccar= 0:03 20 + 18 0:06 + 4 0:1 = travel by bus(Equation)=cbus= 0:03 30 + 0:04 5 + 0:06 3 + 0:1 9 = of choosingmodecar(Equation)=pcarij=e 2:08e 2:08+e 2:18= of choosingmodebus(Equation)=pbusij=e 2:18e 2:08+e 2:18= V , c1ijFigure9:1:logitfunctionTable9:1:Trip characterisitcstuijtwijttijfij :2:Binarylogitmodelexample.

4 Solutiontuijtwijttijfij V , 2007 Table9:3:Tripcharacteristicstvijtwalkijt tijFij ijcoe tripsby car=Tcarij= 5000 2600 Proportionof tripsby bus=Tbusij= 5000 2400 Farecollectedfrombus=Tbusij Fij= 2400 9 = 21600 Whenthefareof busgetsreducedto 6,Costfunctionforbus=cbus= 0:03 30 + 0:04 5 + 0:06 3 + 0:1 6 = of choosingmodebus(Equation)=pbusij=e 1:88e 2:08+e 1:88= tripsby bus=Tbusij= 5000 2750 FarecollectedfromthebusTbusij Fij= 2750 6 = 16500 Theresultsaretabulatedin extendedto a modelcanbe writtenas:P1ij=e c1ij e cmij( ) of tripsfromitojis 5000,andthreemodesareavailablewhich hasthecharacteristicsgivenin Table9:3:Computethetripsmadeby thethreemodesandthefarerequiredto travel by each travel by car(Equation)=ccar= 0:03 20 + 18 0:1 + 4 0:1 = travel by bus(Equation)=cbus= 0:03 30 + 0:04 5 + 0:06 3 + 0:1 6 = travel by train(Equation)=ctrain= 0:03 12 + 0:04 10 + 0:06 2 + 0:1 4 = of choosingmodecar(Equation)pcarij=e 2:8e 2:8+e 1:88+e 1:28= of choosingmodebus(Equationpbusij=e 1:88e 2:8+e 1:88+e 1:28= V , 2007 Table9:4:Multinomiallogitmodelproblem:so lutiontvijtwalkijttijFij ijCeCpijTijcoe :5:TripcharacteristicstvijtwalkijttijFij ijcoe of choosingmodetrain(Equation)=ptrainij=e 1:28e 2:8+e 1:88+e 1:28= tripsby car,Tcarij= 5000 tripsby bus,Tbusij= 5000 , proportionof tripsby train,Ttrainij= 5000 canputallthisin theformofa tableas shownbelow9:4: Farecollectedfromthemodebus=Tbusij Fij= 6 = 9315 Farecollectedfrommodetrain=Ttrainij Fij= 4 = thethirdstageof travel modeis in uencedby erent types of Thetotalnumber of tripsfromzoneito zonejis alternatives-to introducea trainor a characteristicsandrespective coe cientsaregiven in table9 in termsof V , 2007 First,usebinarylogitmodelto ndthetripswhenthereis ,againusebinarylogitmodelto ndthetripswhenthereis travel by car=ccar= 0:05 25 + 0:2 22 + 0:2 6 = travel by bus=cbus= 0.)

5 05 35 + 0:04 8 + 0:07 6 + 0:2 8 = travel by trai=ctrain= 0:05 17 + 0:04 14 + 0:07 5 + 0:2 6 = : Consideringintroductionof bus,Probability of choosingcar,pcarij=e 6:85e 6:85+e 4:09= of choosingbus,pbusij=e 4:09e 6:85+e 4:09= : Consideringintroductionof train,Probability of choosingcarpcarij=e 6:85e 6:85+e 2:96= of choosingtrainptrainij=e 2:96e 6:85+e 2:96= each modeCase1:Tcarij= 4200 4200 :Tcarij= 4200 4200 ,if it is V KrishnaRao